INDEX

Explanations

phrases related to heroic deeds and mentions of specific individuals or titles

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

thinkable

-0.86

ascript

-0.82

accompan

-0.82

picture

-0.81

dropping

-0.80

ãĤ¦ãĤ¹

-0.79

activity

-0.78

advertisement

-0.76

 sugg

-0.75

pattern

-0.75

POSITIVE LOGITS

 Monte

1.09

 Ibid

1.07

 Raven

1.05

Ben

1.05

Ma

1.04

 Rand

1.03

 Kirin

1.03

MN

1.01

Activations Density 1.570%

phrases related to heroic deeds and mentions of specific individuals or titles

No Comments

No Known Activations